The change of carbohydrate antigen 19-9 as a prognostic marker after neoadjuvant chemotherapy in patients with borderline resectable and locally advanced pancreatic cancer

Abstract

Introduction: Although the patients with neoadjuvant chemotherapy (NACT) showed better survival than those who underwent upfront surgery for borderline resectable (BRPC) and locally advanced pancreatic cancer (LAPC), prognostic markers for responders was not confirmed. Method: BRPC (n = 81, 69.2 %) and LAPC (n = 36, 30.8 %) patients who underwent surgery after NACT were included from 2012 to 2017 in tertiary referral center. They were divided into 2 groups based on carbohydrate antigen (CA) 19-9 regression rate (CRR) of 65 % before and after NACT and oncologic outcome was compared. Result: The median pre and post-chemotherapy CA19-9 were 137 U/ml (IQR: 32-371) and 28.7 U/ml (IQR: 13-94.6), respectively. CRR < 65 % (n = 59) showed higher N2 ratio than CRR ≥ 65 % group (n = 58) significantly (20.3 % vs 1.7%, p = 0.021) in perioperative results. Otherwise, there were no significant differences including imaging response (partial response on preoperative computed tomography; 32.2 vs 48.3 %, p = 0.091, maximum standardized uptake value on positron emission tomography; 4.5 vs 3.8, p = 0.310), pathologic tumor regression grade (5.6 vs 11.1 %, p = 0.489) and R1 resection rate (28.8 vs 19.0%, p = 0.279) between CRR < 65 % and CRR ≥ 65 % groups. The patients with CRR ≥ 65 % showed better survival than patients with CRR < 65 % (n = 59) (median survival; 26 vs 47 months, p = 0.024). Multivariate analysis revealed that CRR < 65 % was independent prognostic factor in survival (hazard ratio [HR]; 2.096, p = 0.025, 95% confidence interval [CI]; 1.095-4.009) and recurrence (HR; 1.583, p = 0.053, 95% CI; 0.994-2.523). Conclusion: CRR after NACT can be used for prognostic marker in patients with BRPC and LAPC. However, other markers are needed to evaluate resectability after NACT.